Search Improvements

We realize that there are always further improvements we can make to our search interface. Making these improvements depend on a number of factors: programmer time, involvement in related or associated grants, and how important an improvement is for our users. With those caveats, the following list contains improvements we would like to make in the future.

Searching by date. We are currently evaluating a date normalization tool developed by the California Digital Library (CDL). This tool would hopefully allow us to enable date range searching.
Subject browsing. We have completed a test of browsing records by subject matter using a clustering tool developed at the University of California Irvine (UCI), called the Topic Modeler. Browsing is currently running in our IMLS/DLF Portal. We hope to port what we've learned from this test to OAIster. Additional browsing (i.e., by accessibility, by resource type) remains untested.
Google-like searching. We would like to deploy a single search box that allows Google-like searching, i.e., words entered in the search box are not searched as a phrase, but as single words ANDed together.
Duplicate records. A unique record may be available for harvesting from a number of repositories. We would like to develop a method for providing access to harvested instances of unique records, instead of only harvesting one instance of a record.
Inclusion of thumbnails. One of our grants allowed us to work on including thumbnails in the search results. Currently, thumbnails have been tested and are available in our MODS Portal. Deploying thumbnails on OAIster pends further testing, specifically with the DLF Aquifer Asset Action Packages.

Comments, complaints and suggestions regarding the search interface are welcome. Please send them to Kat Hagedorn (oaister at umich dot edu).

The following is a list of bugs and improvements that were made between 2002 and the present.

Remove instances of HTML from records, except for valid paragraphs and line breaks. Added 2 September 2007.
Truncate display of URL to avoid lengthy horizontal scrolls. Added 21 August 2007.
Remove exact duplicates in format, type, subject, language and year fields. Added 21 August 2007.
Move "weighted hit frequency", our relevancy sort option, to the top of the sort list, so that is the default while searching. Added 21 August 2007.
Change "Keyword" search to "Entire Record" search. Added 14 December 2006.
Allow search limiting by language, in human-readable format. Added 8 March 2006.
SRU compliancy for use with MetaLib and federated search engines. Added 24 February 2006.
Create separate source field populated by DC Source element. Added 7 December 2005.
Ability to save records during a session to a bookbag, and download and email them. Added 30 November 2005.
Add "dataset" as a Type value. Added 27 May 2005.
Make sort pull-down work automatically after choosing a sort option. Fixed 27 May 2005.
Add the Contributor field to the Author/Creator searching. Added 27 May 2005.
Display results counts per institution. Added 15 January 2003.
Ability to search using Boolean operators. Added 31 October 2002.
Fine tuning of title and author sorting. More appropriate insertion of records in the sort that don't contain the field you're sorting on. Creation of relevancy sorting options. Added 31 October 2002.
Contributor field added to records in search results display. Added 31 October 2002.
Ability to revise the search you just made. Added 31 October 2002.
Ability to view all your results, no matter the number. Stopgap measure added 31 October 2002.
Change visted link color of navigation links. Fixed 31 October 2002.
Sort pulldown at bottom of results page not functional. Fixed 31 October 2002.
Choosing weighted hit frequency sorting while searching using truncated words or phrases does not appropriately sort. Fixed 31 October 2002.
Sort order in pull-down menu on search results page haphazard. Fixed 31 October.
Highlight search terms in the search results display for all search fields used. Partially fixed 31 October 2002.
Print 250 characters of "Note" field with link to read rest of field in a pop-up window. User data indicates this is not advisable.
Fix plug.gif problems in menu bar graphics in Netscape 4.x browsers. Not possible to fix; browser idiosyncracies.
Search results labels are not a fixed width. Fixed 2 August 2002.
Multiple instances of URLs are not showing. Fixed 2 August 2002.
Zero results are received when really have received one result. Fixed 2 August 2002.
Should add "all types" to "Resource Type" pull-down on search page. Fixed 2 August 2002.
Language on search results page should be changed so it is easier to determine what fields searched in and which words or phrases were used. Fixed 2 August 2002.
Remove doubled values of "Language" and "Resource Type" fields in search results. Fixed 2 August 2002.
Runtime error occurring on search page in Netscape 4.x browsers. Fixed 1 July 2002.
search.js not being found in Netscape 4.x browsers. Fixed 28 June 2002.
Horizontal rules are not placed appropriately on search results page. Fixed 28 June 2002.
Spacing before first record on search results page is too large in Netscape 4.x browsers. Fixed 28 June 2002.